On the improvement in grey wolf optimization

Grey wolf optimization (GWO) is a recently developed nature-inspired global optimization method which mimics the social behaviour and hunting mechanism of grey wolves. Though the algorithm is very competitive and has been applied to various fields of research, it has poor exploration capability and suffers from local optima stagnation. So, in order to improve the explorative abilities of GWO, an extended version of grey wolf optimization (GWO-E) algorithm is presented. This newly proposed algorithm consists of two modifications: Firstly, it is able to explore new areas in the search space because of diverse positions assigned to the leaders. This helps in increasing the exploration and avoids local optima stagnation problem. Secondly, an opposition-based learning method has been used in the initial half of iterations to provide diversity among the search agents. The proposed approach has been tested on standard benchmarking functions for different population and dimension sizes to prove its effectiveness over other state-of-the-art algorithms. Experimental results show that the GWO-E algorithm performs better than GWO, bat algorithm, bat flower pollinator, chicken swarm optimization, differential evolution, firefly algorithm, flower pollination algorithm (FPA) and grasshopper optimization algorithm. Statistical testing of GWO-E has been done to prove its significance over other popular algorithms. Further, as a real-world application, the GWO-E is used to design non-uniform linear antenna array (LAA) for minimum possible sidelobe level and null control. Performance of GWO-E for the synthesis of LAA is evaluated by considering the several different case studies of LAA that exists in the literature, and the results are compared with the results of other popular meta-heuristic algorithms like genetic algorithm, ant lion algorithm, FPA, cat swarm optimization, GWO and many more. Numerical results further show the superior performance of GWO-E over original GWO and other popular algorithms.